Practical model description of peripheral neural excitation in cochlear implant recipients: 4. Model development at low pulse rates: General model and application to individuals

This fourth paper in the series presents the initial development of the model at low pulse rates, where refractory behavior is minimal. This is a necessary developmental stage that makes possible the subsequent incorporation of temporal effects, in the fifth paper. The model comprises a population of neural fibers spread along the cochlear duct, with normally distributed thresholds. Each has a finite dynamic range, allowing stochastic behavior. The fibers are stimulated by a field that is attenuated longitudinally according to a model-based function, scaled to fit an individual ECAP (compound action potential) measure of effective field attenuation. First, the model parameters were tuned to provide general trend matches to: the observed range of maximum comfortable level (MCL); the relationship between threshold and MCL; and the relationship between curvature of the loudness growth function and MCL. These trend differences between patients are explained by differing percentage of neural survival, although additional parameters clearly influence individual behavior. Second, the model was fitted to three electrodes in each of six subjects implanted with the Nucleus® 24 cochlear implant system (three with straight and three with Contour™ electrode arrays). Across the subjects, different values were required for the “relative spread”, a measure of neural fiber dynamic range relative to threshold, which indicated differences in neural stochasticity. The consistency of the fittings was assessed by comparing model emulations of the ECAP “spread of excitation” (SOE) measure with experimental findings.